An AI with the right file gives the right price and stock, not last month's.
Every page in three forms
Add .md or .json to a tea, brand or story address and you get the same page as clean text or as data. For example, teas.co.uk/ai/model-context-protocol/.md is this kind of story as plain text.
- The web page is for you.
- The text version is for anything that reads words.
- The data version is for anything that wants labelled facts: name, price, stock, barcode.
All three come from the same record, so they can't disagree.
The whole shop in a few downloads
An AI that wants everything doesn't need to visit hundreds of pages. It can download the catalogue in one go:
- products.json and products.jsonl: every product, one record each.
- llm-full.txt: the catalogue as one long text file.
- llm-wiki.txt and llm-recipes.txt: the Tea Wiki and the recipes the same way.
The shop's own rules ask heavy users to take these files instead of thousands of single requests. It's kinder to the shop and quicker for them.
A feed of what changed
catalogue-changes.json lists what changed in the shop, newest first: a price change, a tea going out of stock, a tea coming back, or a product detail updated. An AI can check it and then fetch only what changed, instead of re-reading everything.
It has run since 6 October 2026, and by 20 November its change numbers had reached 2,043.
A tea example. Tea India Ginger Chai changed its stock status at 09:54 on 20 November. That change appeared in the feed straight away. An assistant that checks the feed knows today's answer, not last week's.
What files can't do
A file is only as good as the record behind it. If a fact is wrong in the record, it's wrong in every file. That's why the facts themselves are checked, one by one, on the Truth Spine.
In my words
I used to think of the website as the shop. Now I think of it as one window onto the shop. The shop itself is the records: what we sell, what it costs, whether it's in. Some customers look through the web page. Some look through a text file. Some look through a feed. As long as every window shows the same records, it doesn't matter which one they use. What matters is that none of them goes stale.
Common questions
Why give away the whole catalogue?
Because an AI that can read it can recommend it. Prices and stock aren't secrets: they're what we want people to know.
Are the files free to use?
Yes, under our fair use and citation terms: name teas.co.uk and link the page.
How fresh are they?
The changes feed is updated as things change, and the bulk files are rebuilt whenever the catalogue changes.
Read next: Single source of truth · Everything in one answer · Structured data
Citable formats
For journalists, researchers, AI assistants and content creators. Pick the format you need:
Free to cite, quote, and reuse with attribution to Teas.co.uk.
Got something to add? Logged-in customers can submit additions to the Tea Wiki, admin-approved, your name on the byline, plus reward points.
Sign in to contribute